What "tells" have you noticed for AI-generated text (i.e., typical words/phrases, or other patterns indicative of GenAI)?

2.5k viewscircle icon3 Upvotescircle icon20 Comments
Sort by:
Chief Product Officer in IT Services2 days ago

The inclusion of Markdown ASCII prepending or wrapping of words and phrases

i.e.
# for headings
* for bullet lists
** bold **
etc... etc...

Chief Product Officera month ago

In AI outreach on behalf of direct sales or business development, or even recruiters, it is very clear it is AI-generated when the context is way off - referencing old company associations or completely misunderstanding my role and therefore my interests. There is also over emphasis to "connect emotionally" on something random. Often the text includes hallucinations. The quality is just not there. 

VP of IT in Transportationa month ago

+1 on everything mentioned already. My additional take is when the headers or titles of a paragraph are filled with each word starting with a capital letter.

Lightbulb on1
Director of Product Managementa month ago

Like many have stated, the excessive use of em dashes is definitely the most obvious tell.  On social media posts, I also feel like these are obvious:
 - a list of 3 items with emojis at the beginning
- buzzwordy closing statements (e.g. after all, you want clarity, not chaos)
- never more than 2 sentences together without a space

Lightbulb on1
Program Director, Intelligent Automation + Entrepreneur in Healthcare and Biotecha month ago

The Chatgpt dash! = The em dash

Lightbulb on2

Content you might like

Yes48%

No46%

We’re discussing this possibility5%

View Results

Increased by more than 20%11%

Increased by 10-19%12%

Increased by 1-9%

No change25%

Decreased by 1-9%22%

Decreased by 10-19%25%

Decreased by more than 20%3%

View Results